In WLAN deployments without mesh services, client stations or end stations (STAs) must associate with an access point (AP) in order to gain access to the network. These STAs are dependent on the AP with which they are associated to communicate.
IEEE (Institute of Electrical and Electronics Engineers) standard 802.11s develops a wireless local area network (WLAN) mesh standard. With wireless mesh networks (WMNs) devices can easily interconnect. Each device works as a wireless router that forwards frames for other devices. Thus, networks can be easily deployed without an additional fixed infrastructure.
A so-called mesh network appears functionally equivalent to a broadcast Ethernet from the perspective of other networks and higher layer protocols. As an example, the mesh network may be an LAN according to the IEEE 802.11 specifications, where links and control elements forward frames among the network members. Thus, it normally appears as if all mesh points (MPs) in a mesh are directly connected at the link layer. This functionality is transparent to higher layer protocols.
A standard ‘infrastructure’ wireless local area network is a centralized network in which STAs attach to the AP which acts as a ‘master station’. This centralized topology makes network formation and initial channel selection easy. The AP is configured to start transmitting at a certain frequency channel and the STAs only need to find this channel e.g. by scanning a list of available frequencies. They can do so actively, by broadcasting probe requests on each visited channel, or passively, by listening for advertisements or beacons on each visited channel. After having visited all available channels, they will have found all APs that are in the vicinity, and can select one to associate with.
In many households, a digital subscriber line (DSL) provides high-speed Internet access. Due to economies of scale and strong competition, DSL modems often provide a rich set of features at an affordable price. They do not only integrate an Internet Protocol (IP) router but may furthermore work as print and file server and connect wireless clients via 802.11 links. Accordingly, 802.11 networks have a high penetration rate in the home. Furthermore, APs have become a commodity and can be found nearly everywhere.
Due to the current 802.11 design, the central AP manages the whole WLAN. However, APs typically do not interconnect. Each WLAN established by an AP is an independent network. For large scale coverage, APs require wired backbones that interconnect them. With WMN technology, devices can interconnect over the air. Each device becomes a wireless router that provides the frame forwarding service for other devices. To be able to operate as wireless router, a device needs special capabilities or software modules. However, many existing APs cannot be upgraded. Either the device's manufacturer considers a product to be end of life and thus product maintenance has ended or, the device's hardware limits possible implementations. Thus, a generic solution is needed that connects a WMN with one or more existing APs and thus provides the WMN with the AP's Internet connectivity.
FIG. 1 shows a mesh data frame structure according to the IEEE 802.11 specifications. A frame control (FC) field contains amongst other control information a type and subtype for the mesh data frame and two flags “To DS” and “From DS”. The two flags are set to “1” in order to indicate that the data frame is in the wireless distribution system and therefore in the mesh network. Additionally, address fields A1 to A4 are provided to convey and indicate destination, source, transmitter and receiver addresses. The four address fields contain 48-bit long MAC (Media Access Control) addresses. The first address field A1 indicates a receiver address which defines the mesh point that has to receive the wireless transmission. The second address field A2 indicates a transmitter address which defines the mesh point that sent this wireless data frame. The third address field A3 indicates a destination address which defines the final (layer 2 or link layer) destination of this data frame. The fourth address field A4 indicates a source address which defines the (layer 2 or link layer) source of this data frame.
Furthermore, duration/identity (D/ID), sequence control (SC) and frame check sequence (FCS) fields are provided, which are not discussed here for brevity and simplicity reasons. Further details can be gathered from the IEEE 802.11 specifications. A body (B) portion is provided to convey desired payload data up to a length of 2304 octets.
Each of the above addresses may have a length of 6 octets and maps on the address fields A1 to A4 in dependence on the “To DS” and “From DS” information of the FC field. The IEEE 802.11 standard clearly states that an address field is omitted “where the content of a field is shown as not applicable (N/A).” Solely when both bits “To DS” and “From DS” are set to “1”, four address fields appear in an 802.11 frame.
FIG. 2 shows a schematic network architecture, where the four address fields A1 to A4 are used to interconnect to different IEEE 802.3 network segments with the help of a wireless network (e.g. an IEEE 802.11 WLAN) comprising devices G and H. Here, the wireless network is used to provide a bridge between a first independent wired LAN comprising devices A to C and a second independent wired LAN comprising devices D to F.
APs form infrastructure basic service sets (BSSs). In a BSS, the AP relays all traffic. Although the IEEE 802.11 standard provides four address fields, only three address fields A1 to A3 are typically needed in an infrastructure BSS.
FIG. 3 shows a signaling example based on the network architecture of FIG. 2, wherein device B sends a frame destined to device F. In this case, four address fields are needed on the wireless link between devices G and H. The fourth address field A4 corresponds to a source address (SA) field that holds device B's address. The third address field A3 corresponds to a transmitter address (TA) field that holds device G's address. The first address field A1 corresponds to a receiver address (RA) field that holds device H's address. The second address field A2 corresponds to a destination address (DA) field that holds device F's address. Once device H has successfully received the 802.11 frame from device G, it strips off the data portion in the 802.11 body and sends out the data portion in an 802.3 frame that solely contains device B's address as source address and device's F address as destination address.
However, most of the current 802.11 APs, however, cannot operate in this bridging mode as described above. They solely serve as AP in their local infrastructure BSS.
FIG. 4 shows an exemplary conventional network architecture, where a single device C works as AP, router and modem that connects a WLAN with an external network 100, e.g. the Internet. Client devices A and B have associated with the AP C. The devices A and B may exchange data via the AP C or access the external network 100. In any case, the devices A, B and, C use three address fields only. The client devices A and B set the “To DS” bit to “1” and the “From DS” bit to “0” when sending a frame to the AP C. If the device B sends a frame to the device A, the first address field A1 contains the AP's address as the receiver address (RA). The second address field A2 contains the source's address. Here, the source address is device B's address. The third address field A3 contains the ultimate destination's address. Here, the destination address (DA) equals device A's address. In case device B wants to communicate with an Internet station, the DA address holds the AP's address since device C works as an IP router or default gateway.
Once AP C has received device B's frame, it analyzes the third address field for the destination address. If destined to AP C, the frame is send to a higher layer where the IP router operates. If destined to device A, AP C relays the frame. Thus, AP C sends a frame to device A that has the “From DS” bit set to one and the “To DS” bit set to zero. The frame's address field 1 contains the Receiver Address (RA). In this case, it contains device A's address. The second address field contains the Transmitter Address (TA) that is AP C's address. The third address field contains the Source Address (SA), which equals device B's address.
FIG. 5 shows a correspondingly reduced 802.11 mesh data frame structure with only three address fields A1 to A3, sufficient for all of the aforementioned frames in an infrastructure BSS that uses three address fields only.
With APs that can handle three addresses only, the local BSS is limited to a single wireless hop. FIG. 6 shows an exemplary conventional network structure, where a device A connects with an AP C and a WMN. The device A can use any service of the AP C's infrastructure BSS. However, the device A cannot provide its network connectivity to the WMN. Although the device A and other devices D to G form a single WMN, the addressing limitation at the AP C prevents the device A from sharing its connectivity.
Since the AP C allows for the usage of three addresses only, all frames transmitted to the WMN must be destined to the device A. Without further information however, the device A cannot decide about a frame's final destination. Thus, the device A cannot forward frames to another destination in the WMN.